The present paper uses a data-driven research paradigm to conduct a systematic examination of the distribution channels of Chinese traditional musical instruments around the world music scene. Initially, it will be done through acoustic analysis of audio samples by extracting and refining audio features according to the MPEG-7 standard (spectral centroid, spread spectrum, and their first- and second-order derivatives) using the K-Nearest Neighbors (KNN) classification algorithm to recognize instruments automatically. On top of this base, a retrieval system has been built to reach efficient retrieval of target instrument sample out of global music databases. Another concept of agent-based modeling was presented and the classes of agents were defined: propagators, receivers, and immune agents with their interaction rules and the Runge-Kutta method was used to perform numerical simulation. The results indicate that adding dynamic eigenvalues is much useful in improving classification performance where the mean F1 score increases by 86.29% to 96.42%. Moreover, experimental harmonic structure analysis determined the best harmonic order to be 7, which increased the recognition rate of all four categories of instruments to over 93.Finally, to simulate the propagation mechanism, the equilibrium stability of musical information transmission was revealed: when the basic reproduction number R₀ = 3.5, the system stabilizes after approximately 8 days. The study further demonstrated the significant impact of knowledge agents’ individual characteristics across different network layers (In, Middle, Out) on knowledge diffusion rates and system evolutionary dynamics. This research provides novel insights and tools for promoting Chinese traditional musical instruments globally.
